Pere Fàbregas Salis 
Unpublished Conjectures by Nicolaus Heinsius on Ovid’s Metamorphoses 1–4 

Abstract: This paper publishes for the first time 132 conjectures by Nicolaus Heinsius on 
Ovid’s Metamorphoses 1‒4. The value and possible motivations of each proposal are briefly 
assessed. 

Keywords: Ovid; Metamorphoses; Nicolaus Heinsius; unpublished conjectures; textual 
criticism 

Nicolaus Heinsius accompanied his colossal 1659 edition of Ovid’s Met. with a bulk of 
more than 400 pages of critical notes,1 which he, constricted by public duties, regretted 
having composed in a hurry.2 Later, using Heinsius’s manuscript drafts, Burman (1727) 
was able to supplement Heinsius’s original notes and to produce his own. After that, 
Heinsius’s rich preparatory materials, which included not only collations of many 
manuscripts, but also unpublished conjectures, fell into almost complete oblivion.3 The 
goal of this paper is to publish conjectures for books 1‒4 present mostly in the 
Berolinensis Diez. 4º 1075 (= b3)4 which have never been printed by Heinsius himself 
(1659) nor Burman (1727), Magnus (1914) or Slater (1927),5 who made some use of 
Heinsian drafts.6 I comment briefly on each proposal.7 
It is perhaps worth saying that I have decided to include every suggestion, not only the 
strongest ones, because I might have misjudged them and because, at this point, any 
exclusion would amount to a second damnatio memoriae. Furthermore, even what one 
might call coniecturae nugatoriae8 have something to teach us. On the one hand, they 
reveal extraordinary knowledge of the mechanisms of transmission, and some certainly 
have at least diagnostic value.9 Otherwise, by provoking refutation, they advance our 

1 This is his second edition. The first one of 1652 has no commentary. 
2 Heinsius (1658) 6r–v* (Nicolaus Heinsius Candido Lectori S.). Heinsius was indeed heavily busy with 
diplomatic affairs up to 1671; similar regrets also appear in other editions of his up to that year. An 
excellent, though incomplete, account of Heinsius’s career is still Blok (1949). For an updated critical and 
biographical profile of Heinsius, especially as regards his travels to Italy, see Celato (2023). 
3 A brief account of the fate of these materials is to be found, e.g., in Fàbregas (2022) 35‒37. 
4 On the Berolinensis, see Boese (1962) 160‒161 and 170; Reeve (1974) 136 n. 9 and 150‒151; Winter 
(1986) 16‒17; Fàbregas (2022) 36 n. 12. 
5 Nor in Heinsius (1652, 1742, 1758) or by Suárez (2015) in the case of book 3. Additionally, since the 
proposal has been neglected, I point out that Heinsius’s ingentesque animo et (b3, 25r) at 1.166 was 
communicated per litt. to Gronovius (ap. Burman 1727 [Syll.], 293 [cf. 292]); however, he must have 
later realized that concipit is actually introduced by et in 164 (cf. Burman 1759, 126). Likewise, Ramírez 
de Verger (2022) 89‒90 has recently attributed to Heinsius (b3, 58r) 2.382 quali, cum deficit, orbem 
(Magnus 1914, 66 assigned it to Passerat), probably an attempt to make sense of the reading orbe. 
6 Munari (1950) 164‒165 published some conjectures on the Amores from the Oxonienses Bodl. Auct. S 
V 7 and Auct. S V 10. Despite these rather isolated efforts, a comprehensive exploration of Heinsian 
manuscript notes has not yet been undertaken. For my part, I intend to at least partially mend this state of 
affairs and gradually disclose all his unpublished conjectures on Met. 
7 The lemmata are from Tarrant’s OCT (2004). I indicate in brackets the exact folium of b3 in which they 
appear. Furthermore, the following sigla are also used for Heinsius’s libri impressi cum notis 
manuscriptis: b2 = Berolinensis Diez. 4º 1072; o4 = Oxoniensis Bodl. Auct. S V 8. When Heinsius quotes 
ms. readings, I try to identify them to the best of my ability using the sigla available at 
http://www.uhu.es/proyectovidio/ing/index.html (a version of the website is stored for long-term 
preservation by the Bayerische Staatsbibliothek: https://opacplus.bsb-
muenchen.de/search?oclcno=973043726&db=100&View=default). The sigla are also available in Díez 
(2014); Suárez (2015); Rivero (2018); Ramírez de Verger (2021); Fàbregas (2022). 
8 Heinsius himself defined so some of his proposals added in the margins of his collation of the codex 
Etruscus of Seneca’s tragedies in a letter to Gronovius (ap. Burman 1727 [Syll.], 203). 
9 Cf. Maas (1960) 32. 

This is an accepted manuscript (AM) of an article published by De Gruyter in
Philologus : Zeitschrift für antike Literatur und ihre Rezeption = A Journal for Ancient Literature and its Reception, vol. 168, no. 1, 2024, pp. 42-69, available at https://www.degruyter.com/  
https://doi.org/10.1515/phil-2023-0024.
It is deposited under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial 
re-use, distribution, and reproduction in any medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way. If you wish to use this manuscript for 
commercial purposes, please contact rights@degruyter.com.



 
 

understanding of the passage at hand (Maas 1960, 13). In the worst-case scenario, they 
allow us glimpses into the officina of the uerus sospitator Ouidii. 
 
1.15‒16 utque erat et tellus ... / sic erat] quaque errat tellus (uel quaque haeret tellus) ... / 
siderat (17v) 
 
In addition to quoque ierat tellus (ap. Burman 1727, 13 = Heinsius 1758, 427),10 which 
makes good sense and avoids the repetition of erat, Heinsius made other conjectures, 
also implying the notion of movement: quaque errat tellus, drawing inspiration mostly 
from quaque erat et tellus of “optimus Cantabrigiensis [deperditus]11 & primus 
Palatinus [sc. Gruteri = Hd; also V15 V26]” and quaque erat tellus of “2 Pall.” (V9 
V17); and quaque haeret tellus. He even ventured siderat instead of sic erat, which, 
despite being palaeographically excellent, does not seem compelling to me. We should 
keep the combination ut ... sic (cf., e.g., 1.45‒47; 1.370; 4.440‒441). Read Burman ad 
loc. The reading qua instead of ut is probably a gloss. 
 
1.92 turba] torua (21v)12 
 
The “Langermanni codex” (Gf42m(n.l.); also Bo) had uerba, which just probably slipped 
from the previous line. However, Heinsius, as reported by Burman (1727, 23 = Heinsius 
1758, 430), took the chance to suggest terra, which seems flat to me (this could even be 
Burman’s mistake, since I cannot find the proposal among Heinsius’s drafts). Much 
more interesting is Heinsius’s autograph proposal torua, which is palaeographically 
easy and provides a more pointed sense than turba. Cf. Petron. Sat. 123 vers. 206 toruo 
Iuppiter ore; Sil. Pun. 12.551‒552 stat celsus et asper ab ira / ingentemque metum 
toruo domat ore senatus; Stat. Silv. 5.2.179; Achil. 2.52‒53. 
 
1.155 subiectae Pelion Ossae] subuecto Pelio Ossae (23v) 
 
Jupiter knocks off the ladder to the heavens the Giants built piling up mounts Olympus, 
Ossa and Pelion. Since Homer (Od. 11.315‒316), Pelion was often on the top, and Ovid 
kept that order in Am. 2.1.13‒14 (with McKeown 1998, 13) and Fast. 1.307‒308; 
3.441‒442; cf. also Prop. 2.1.19‒20. Thus, subiectus must be probably referred to 
Ossa.13 Nevertheless, maybe because Virgil (Georg. 1.281‒282) had reversed the 
Homeric order, Heinsius was also prepared to read subiecto Pelio Ossam, which he had 
found “in uno Ambrosiano pro diversa itidem lectione” (A2).14 But probably because 
there seems to be no reason why Ovid would resort to the Virgilian order here, in b3 he 
suggested subuecto, which would keep Pelion on top; but I cannot find fully comparable 
examples of this alleged used of subueho. 

 
10 See also Burman (1727 [Syll.]) 395. 
11 Vid. Reeve (1974) 150 n. 67; Fàbregas (2019) 86. 
12 Also o4, 5. 
13 Perhaps regarding the dative with excutere as objectionable (the construction is certainly not common 
[vid. OLD s.v.], and the thought seems a bit odd), Heinsius recommended (1659) 13‒14n subiectam Pelio 
Ossam, which is excellent. 
14 On f. 2v, A21 reads subiectum pelion oss(a)e, A22v added ossam and A23m pelio, but, as far as I can tell, 
there is no correction on subiectum. A further argument for this interpretation is that the first mountain 
struck by lightning is Olympus, which we could assume to be placed on top. Nevertheless, excutere rather 
suggests that Jupiter struck the lowest mountain first and the others were “shaken off”. 



 
 

 
1.173 hac parte] at fronte (24v) 
 
I am not very convinced by the reading of Be hac parte, which Bömer (1969) 79 rightly 
defines as colourless. The vulgate a fronte or hac fronte would place the godly halls in a 
clearly conspicuous place on both sides of the Via Lactea and close to the top, 
contrasting with the dispersion of the ‘plebeian’ houses. However, a or hac are still dull 
and somewhat prosaic. Heinsius thought of at, which is great. At fronte could have 
easily been corrupted to hac fronte (perhaps because of hic in 175) or to the more 
common expression a fronte. This conjecture deserves serious consideration. 
 
1.175 uerbis] ueris (24v) 
 
Heinsius cleverly suggested ueris, which could have been replaced by an obvious gloss, 
palaeographically very close, such as uerbis. This variation is attested at 10.20 (vid. 
Fàbregas 2022, 118). But cf., e.g., Cic. De or. 3.36 alterum enim exultantem uerborum 
audacia reprimebat; Orat. 81.28 nec in faciundis uerbis erit audax. 
 
1.206 compressit] compescit (26v) 
 
Heinsius probably wanted to avoid the repetition compressit / pressit (207) in a rather 
tautological sequence. He quoted 14.629‒630 spatiantia passim / bracchia compescit; 
Tr. 1.2.87 tantos compescite fluctus; 5.6.45 intempestiuos igitur compesce tumores. To 
these, add Sen. Herc. f. 47.3 uirga murmur omne compescitur. Cf. also Claud. rapt. 
Pros. 1.84‒88. Compescit is certainly attractive. However, cf. Epiced. Drusi 467 iam 
comprime fletus; Stat. Theb. 1.207 compressa metu seruantes murmura Venti; and 
perhaps also Lucr. 1.69 murmure compressit caelum.15 Vid. Bömer (1969) 90 ad loc. 
 
1.280 immittite] admittite (26v) 
 
Permittite of “unus meus [P9], quod magis placet” (b3, 30r) would be indeed possible 
(cf. ThlL VI.3 2394.57‒58). Less likely is dimittite of “tres alii” (e.g. Hd Lr19P24 Lr26; 
cf. Stat. Theb. 9.233 longum dimittere habenas; August. Conf. 8.12.28) and even less so 
emittite of “unus” (Bo2; there seems to be no classical parallels for habenas emittere; 
one would expect flumina as object). Being aware that prefixes are often wrong in this 
paradosis (cf. Luck 2008, 11‒12), Heinsius added “posset et admittite legi”, which 
could certainly be right: cf. 11.512 ubi se uentis admiserat unda coortis; Val. Fl. 6.303 
opprimit admissis ferus hinc Gesander habenis. Nonetheless, as recalled by Bömer 
1969 (107), immittere habenas is a common poetical expression (cf. ThlL VI.3 
2394.49‒51; OLD s.v. immittere 9a). 
 
1.298 uineta] pineta (26v) 
 
Heinsius commented: “malim et pineta. pinus enim inter proceras arbores”.16 This could 
certainly go well with the immediately preceding lines, which seem to imply fairly deep 
waters (295‒297); but nothing rules out uineta (cf. 292‒295). 

 
15 In strictly physical sense, cf. Ov. Met. 6.556; also 6.294 (with Ramírez de Verger 2021, 170‒174). 
16 A similar confusion is attested at 11.87 (uineta/pineta/spineta: vid. Díez 2014, 69‒71) and 6.15 
(uineta/dumeta: vid. Ramírez de Verger 2021, 72). 



 
 

 
1.406 non] nondum (33r) 
 
Heinsius suggested nondum exacta satis. Nondum satis is mainly prosaic, but not rare. 
The first poetic example I can find is Luc. 8.786‒787 semusta rapit resolutaque nondum 
/ ossa satis neruis. Cf. also Ov. Her. 17.177‒178 nec adhuc exacta uoluntas / est satis. 
Since the sense would remain intact, -dum could have dropped in synaloepha. Cf. 
Fàbregas (2022) 438. For the synaloepha, cf., e.g., Met. 9.17. This is indeed an 
interesting proposal. 
 
1.440 tantum spatii] totumque spati (34r) 
 
Heinsius attested: “tantumque spatii unus Med. ex melioribus [Lr7]”. Since the reading 
is unmetrical, he suggested totumque spati. This phrase explains why Python was so 
terrifying, which makes the asyndeton preferable. Furthermore, totum does not seem to 
admit the partitive construction (cf. Pinkster 2015, 991–992; cf. OLD s.v. 4b; one would 
rather expect totumque spatium). Likewise, the contracted form spati should be avoided 
(vid. Neue/Wagener 1902, 146; Bednara 1906, 340‒341; Fàbregas 2022, 278‒279 on 
10.359). 
 
1.533 ut] ceu (36r) 
 
The variation between ut and ceu is documented at 11.26 (vid. Díez 2014, 16‒17).17 For 
ceu ... sic, cf., e.g., Verg. Aen. 10.723‒729. Vid. ThlL III 982.5‒9. But there is no need 
for the change; for ut ... sic, see on 1.15‒16.18 
 
1.542 sparsum] passum (36v) 
 
This is again a very interesting proposal. Sparsum could be an easy gloss of the much 
more difficult passum. The confusion between the two participles is almost systematic 
in the mss. (in similar syntagms, e.g., Met. 2.238 passis ... comis; Tr. 1.1.12 [Enn. Ann. 
349 Skutsch]; Met. 4.521 passis ... capillis; 5.513; 6.531 [see Ramírez de Verger 2021, 
249]; 7.257; 11.49; Her. 6.89; Ars am. 3.709; Fast. 2.813; Tr. 1.3.43; Am. 3.2.45 passis 
... pinnis [with Heinsius 1658, 267n]). For crinis passus, cf., e.g., Verg. Aen. 1.480; 
2.403‒404;19 Ov. Fast. 1.645 (v.l. sparsos); Tr. 4.2.43 (v.l. sparsis). Vid. ThlL IV 
1204.35‒38; X.1 195.67‒196.2. However, as far as I know, the juncture always appears 
in the plural. On the other hand, Ovid does not use crinis sparsus anywhere else (but cf., 
e.g., Sen. Herc. f. 473; Oed. 416; Phaedr. 393; Luc. 9.683; vid. ThlL IV 1204.53‒58). 
See also on 4.261. 
 
1.569 claudit silua] claudunt saxa (uel lustra) (37v) 
 
Heinsius attested: “silva [Lr9], vel, littora unus Mediceus [Lr92m]”. From this, he 
suggested “forte claudunt saxa vel, lustra” (in o4, 21, he added ora). Tempe is a defile 
between mounts Olympus and Ossa (OCD s.v. Tempe) which is often described as 
woody and shady. Lustra could certainly match this description, and the word could 
have been replaced by its gloss silua (could littora be a corruption of lustra?). Saxa 

 
17 Heinsius also pointed out that at 1.555 one of his mss. (Gf72) had ceu instead of ut. 
18 The simile is modelled upon Verg. Aen. 12.749‒757, where ueluti is used as introduction. 
19 Some recc. of Virgil have passis (see Burman 1746, II, 132 and 264). 



 
 

seems less likely, ora even less so. In any case, the transmitted silua looks fine: cf. 
Catull. 64.285‒286 confestim Penios adest uiridantia Tempe, / Tempe quae siluae 
cingunt super impendentes (with Ellis 1889, 327); Plin. NH 4.31. 
 
1.569‒570 ab imo ... Pindo] ad imum ... Pindum (37v) 
 
Heinsius’s suggestion cannot be right. The river Peneus flowed from mount Pindus into 
the sea through Tempe. Bömer (1969) 179 quotes Strabo 7 frg. 14 (ῥεῖ δ’ ὁ Πηνειὸς ἐκ 
τοῦ Πίνδου ὄρους διὰ μέσης τῆς Θετταλίας πρὸς ἕω) and 15 (ὁ Πηνειὸς ποταμός, ῥέων 
διὰ τῶν Τεμπῶν, καὶ ἀρχόμενος ἀπὸ τοῦ Πίνδου ὄρους). See also Anderson (1997) 
203‒204. Therefore, ab imo ... Pindo is right. 
 
1.591 altorum nemorum] dumorum aut nemorum uel horum altas nemorum (39r) 
 
The transmission of this line is seriously troubled. Heinsius wanted to avoid the 
repetition of altus (592 altissimus). In addition to his printed suggestions (1659, 25n; cf. 
Burman 1727, 73), after dū horum nemorum aut horum of the “sec. Palat.” (Hd1), he 
conjectured dumorum aut nemorum, which is palaeographically clever, but makes the 
repetition in parenthesis clumsy.20 He also attempted horum altas nemorum, which is 
better, but does not remove the repetition. 
 
1.737 hoc audire] exaudire (44r) 
 
Heinsius wrote: “haec audire Cantabr. [deperditus]21 VI. alii [e.g. Lr2 Lr3 Bs2 Gf8Lr14 
Lr20]. forte, exaudire”. Just after the direct speech, hoc might seem a bit weak. 
Exaudire provides an apparently pointed sense: cf. Am. 2.9.51.22 But it often entails the 
notion of “heeding” (OLD s.v. 3b), whereas the Styx seems here a mere witness to 
Jupiter’s tricky promise. Cf., e.g., 6.548; Verg. Aen. 12.200. 
 
1.771 ipse] ecce (44v) 
 
Esse of “sec. Pal.” (Hd) prompted the conjecture. The confusion between esse and ecce 
is possible (cf. 2.78), but ecce is not very attractive here. The reading of Hd probably 
just tried to make clear that uidendum must be understood as uidendum esse. On the 
other hand, there is nothing wrong with ipse (cf., e.g., Lucr. 3.874‒875 quamuis neget 
ipse / credere se quemquam sibi sensum in morte futurum; Ov. Met. 9.752 non se negat 
ipsa roganti), even if one might prefer the reading ille. 
 
2.125 repetens] promens (50r) 
 
The conjecture seems to be inspired by reprimens of N4 (repetens N4v), but it is unlikely 
that N4 is singling out a corruption, nor do I have reason to believe that the transmitted 

 
20 On the repetition in “demonstrative parenthesis”, see Wills (1996) 335‒341, esp. 339. 
21 See n. 11. 
22 For a similar corruption, vid. Ramírez de Verger (2021) 172‒173 on 6.294. The preceding haec might 
have encouraged the confusion. 



 
 

repetens is an interpolation from 13.739 (with Rivero 2018, 331).23 Furthermore, I have 
not found parallels for promere suspiria. 
 
2.127 et fortius] at fortiter uel sed precor (50r) 
 
Burman (1727) 102 claimed credit for at fortiter, which is attractive, but it should be 
attributed to Heinsius.24 As an alternative, he also suggested sed precor, which is dull 
and far too removed from the paradosis. 
 
2.138 tortum ... Anguem] tortam ... Anguem (50r) 
 
As the “primus Moreti” (unidentified) read inter utramque in 140 (by assimilation with 
pressam ... ad Aram?), Heinsius suggested tortam ... Anguem here. It is true that scribes 
tended to restore the masculine with anguis,25 but the constellation is regularly 
masculine (cf. ThlL II 55.46‒55). 
 
2.146 est] si (50v) 
 
Heinsius commented: “sit unus Leid. [Ld21] Lego si non invenusta repetitione”. But the 
repetition and resulting thought has no true rhetorical sharpness. The enjambement of 
est tibi looks far more acute. Cf. Fast. 4.600. 
 
2.277 subsedit] subsidit (53v) 
 
Lines 275‒279 are all in the perfect. Heinsius’s subsidit must be therefore wrong. Vid. 
Fàbregas (2022) 248 on 10.284. 
 
2.316 illic ... illic] istic ... illic (54v) 
 
O4 reads istinc ... illinc, which prompted Heinsius’s istic ... illic (no examples in Ovid 
or in classical literature: cf. ThlL VII.1 373.52). See Bömer (1969) 321; Anderson 
(1997) 263. 
 
2.399 dolens] premens (59v) 
 
Unless deeper corruption should be supposed, dolens was misunderstood once taken 
with stimulo and uerbere, giving way to glosses (furens) or corrections (domans). After 
the unlikely fremens of “unus Leid.” (Ld8, f. 16r, reads tremens, on which Heinsius [b2, 
50v] commented “l(ege) fremens”), Heinsius conjectured premens, which is possible, 
but inferior to dolens: cf. 11.345. 
 
2.400 enim] equis (60r) 
 

 
23 For the simplex in a similar context, cf. 2.621‒622. By the way, whether repetens means that Phoebus 
repeatedly sighs while he talks (cf. Lygd. 6.61) or whether this a case of “vi praepositionis evanida” 
(Bömer 1969, 272‒273; cf. Bach 1831, 68), I cannot tell. 
24 Heinsius already observed that “primus Med. [Lr] unus Bonon. [Bo3]” had et fortiter. One of the 
anonymous reviewers points out that in his Adversaria Heinsius (1742) 160 quoted the line with the 
reading et fortiter. A similar oscillation between utilius and utiliter is attested at Am. 1.7.24. 
25 Cf. 4.495‒496 (with Heinsius 1659, 95n; Burman 1727, 284); 4.594 (with Heinsius ap. Burman 1727, 
293). 



 
 

Heinsius (1659) 37n suspected lines 398‒400 (400 imprimis, deleted by Tarrant 2004, 
46). Probably for this reason, he suggested equis instead of enim, but the repetition of 
the word is not great (especially with its reprisal with illis at the end of the line). For the 
parenthesis with enim, cf., e.g., 15.566. 
 
2.404‒405 terras hominumque labores / perspicit] terrisque hominumque labori / 
prospicit (59v)26 
 
Heinsius certainly found inspiration in prospicit of Lr P24 Lr20 (also Go2).27 The 
confusion between the two verbs is common enough (ThlL X.2 2219.12‒14). The 
enclitic after terris seems out of order (he excludes it in o4, 41), but, as Jupiter is 
restoring cosmic order after Phaethon’s race, prospicit with the dative could certainly be 
right. Available examples of this construction are mostly prosaic (vid. OLD s.v. 6; ThlL 
X.2 2222.15‒60), but cf. 15.834‒836 inque futuri / temporis aetatem uenturorumque 
nepotum / prospiciens. Despite this, perspicio seems faultless (vid. Bömer 1969, 269 on 
2.112).28 
 
2.424 rescierit] rem scierit (61r) 
 
In addition to rescieris attested by Slater (1927) ad loc.,29 maybe because rescisco is 
very rare in Augustan poetry (see Bömer 1969, 347‒348: “bei Ovid nur hier”), Heinsius 
also suggested rem scierit, which is palaeographically obvious. In b3, Heinsius does not 
mention that this is the reading of Bo3 (also Li3). But cf., e.g., Plaut. Asin. 743 ne uxor 
resciscat metuit; Ter. Phorm. 585 uereorque ne uxor aliqua hoc resciscat mea. 
Furthermore, the use of both compounds in close proximity (nesciet/rescierit) seems 
deliberate. 
 
2.424 sunt o sunt] quin sunt haec (61r) 
 
Trying to make sense of quia sunt haec of “alius”,30 Heinsius attempted quin sunt haec, 
which gives excellent sense. But for a similar pattern to sunt o sunt, cf. 4.583 accede, o 
coniunx, accede; Pont. 1.1.59‒60 paenitet, o ... / paenitet; Verg. Aen. 8.579 nunc, o 
nunc liceat crudelem abrumpere uitam (cf. Stat. Theb. 3.360). 
 
2.451 et] ac (61r) 
 
The “pr. Medic.” (Lr) reads at, which is wrong. Hence Heinsius suggested ac. But et 
and at are regularly confused.31 
 
2.458 est] en (62r) 
 
Diana tells Callisto and her other companions that they are alone (so it is safe to take off 
their clothes and bathe). Heinsius’s en is attractive.32 Cf., e.g., 8.590‒591 ut tamen ipse 

 
26 Also o4, 41 with terris instead of terrisque. 
27 Also Douza (teste Burmanno 1727, 124) by conjecture. 
28 Bach (1831) 86 observes that perspicit matches explorat (403). 
29 Reading sunt o tua with P2. 
30 The closest reading I have been able to trace is quid sunt o of Lr22. 
31 On the variation, see Rivero (2018) 413; Fàbregas (2022) 277‒278 and 431. 
32 En calls attention to something perceptible, which can also be a situation (Pinkster 2021, 931). In any 
case, on Heinsius’s taste for en, see Rivero (2018) 164 on 13.262. 



 
 

uides, procul, en procul una recessit / insula, grata mihi. The confusion between en and 
est is palaeographically easy.33 
 
2.473 mei] mihi (62r) 
 
The dative is probably more poignant than mei,34 but cf. 475 places nostro, importuna, 
marito; Fast. 6.52 hic teneam cum Ioue templa meo. 
 
2.482 precantia] rogantia (62r) 
 
Heinsius merely wrote: “l(ege) rogan-”. I had thought that Heinsius independently 
predated Schepper (ap. Faber 1665, 329), but his source was likely Gronovius, who 
communicated the proposal to him in a letter of the 14th January 1658.35 
 
2.494 conspectos] consuetos (63r) 
 
The conjecture is no improvement. In fact, conspectos is great: cf., e.g., Her. 19.205‒
206 ut te conspecta subito ... / restiterim. 
 
2.518 †est uero quisquam†] estne igitur cur quis qui (62v), est autem ut quisquam (uel 
est autem cur quis) (63r) 
 
Instead of ergo in his estne ergo quis qui (1659, 39n), Heinsius also suggested igitur, 
which is possible, but palaeographically one step further away.36 Additionally, he 
attempted estne autem ut quisquam or est autem cur quis. The latter is tolerable (cf. 
13.114; Her. 10.144; Pont. 1.7.54), but I do not make much sense of the former. 
 
2.559 Pandrosos] Pandrasis (65r) 
 
Heinsius attempted Pandrasis, probably after Pandrasos of most mss., and referred to 
2.738 (where similar variants are recorded).37 Perhaps Pandrosis would be feasible, yet 
this seems a second declension noun (cf., e.g., Paus. 1.27.2‒3; read Bömer 1969, 375 
and 376‒377). Thus, Pandrosos must be right (vid. Sniehotta 1903, 5‒6).38 
 
2.563 dicar] cedar (64v) 
 

 
33 Vid. Heinsius (1659) 426n on 15.677. 
34 Mei might be taken as a sign of affection (cf. Anderson 1997, 288). Heinsius apparently had a certain 
penchant for the dative: see on 3.191; 3.456; 4.287; and cf. Fàbregas (2022) 338 and 390. 
35 Ap. Burman (1727 [Syll.]) 382. On the 24th December, Heinsius (ap. Burman 1727 [Syll.], 377) had 
confessed to Gronovius that he was struggling with the line and that he propended for neue truces animos 
per uerba precantia flectat (cf. 1659, 38n). I thank one of the anonymous readers for the precision. 
36 The confusion is rather common (ThlL VII.1 254.66‒69; V.2 762.8‒9): cf. Cic. Div. Caec. 19 quis ergo 
[igitur v.l. teste Peterson] est qui neget; Planc. 24 quis est igitur [ergo dett. aliquot teste Olechowska], qui 
neget. For igitur, cf. Sull. 52 num quis est igitur, qui tum dicat; Orat. 225 quis est igitur qui non fateatur. 
37 Cf. Tarrant (2004) 491. 
38 At 2.738 Heinsius (1659) 55 (also b3, 72r) read Pandroso instead of Pandrose. But, as a second 
declension noun, the vocative in -e is certainly expected; if this was originally a noun in -ω, -ō would be 
required in Latin (vid. Neue/Wagener 1902, 449; Sniehotta 1903, 15‒16). Similarly, at 4.45 Heinsius 
(1659) 79n wrongly wanted to replace the vocative Dercetĭ for Dercetŏ. Cf. Bömer (1976) 29. 



 
 

A “Gronovianus” (B2) had credar,39 whence Heinsius suggested cedar,40 anticipating 
Watt’s (1999) 167‒168 cedam. The verb fits the context and balances well ponar in the 
next line. The passive (vid. OLD s.v. 13) is palaeographically more adequate and better 
demonstrates the crow’s reluctance at being replaced as Athena’s protégée. But for 
dicar, cf. Housman (1920) 40 on Manil. 4.314. 
 
2.566 nec] uel (64v) 
 
The crow argues that he had not instigated Athena to seek him out. Heinsius’s uel is 
possible, but not really an improvement. 
 
2.568 negabit] negarit (65r) 
 
Heinsius noted: “negabat unus [Bo3]. Lego negarit.” However, the future seems better 
after quaeras and irata est. 
 
2.586 retinebat] retinebar (66r) 
 
I would happily endorse Heinsius’s suggestion, if only I had been able to locate 
parallels for retineo with the Greek accusative.41 
 
2.589 prodest] refert (66r) 
 
Heinsius quoted 13.268 Quid tamen hoc refert, si..., where the gloss prodest has taken 
the place of refert in some recc. (vid. Rivero 2018, 167). Heinsius could well be right. 
 
2.591 an] num (66r)42 
 
Heinsius’s num is based on nunc of “duo libri” (Bo2, an Bo22v).43 But an is said to 
imply a certain disbelief of the speaker, which fits the context best. 
 
2.603 arma adsueta capit] armaque sueta rapit (67r) 
 
This proposal derives from sueta (“bene”) instead of adsueta of three mss. (L7 Lr141 
O12).44 However, ad- was probably just lost in synaloepha.45 Furthermore, Ovid never 
uses suetus. 
 
2.637 in rapidi] Aemathii (68r) 
 

 
39 Credar is probably a gloss of dicar. 
40 Apparently, Heinsius also suggested replacing at for ut (a “Regius”, which I have been unable to trace, 
reportedly had ad dicar). We should then punctuate after redditur. But the change seems pointless. 
41 Even if it is a different example, cf. 2.577. 
42 Also o4, 47. 
43 In o4, 47, Heinsius assigns nunc to Lr10 and Bo2; Lr10 has nunc instead of non in 592. I suspect 
Heinsius inadvertently mixed collations. 
44 Heinsius (1659) 51 and 41n printed rapit of “Mediceus unus ex melioribus” (Lr22). Cf. Verg. Aen. 
10.462. 
45 Unless one imagines that the scansion sŭētă (Lucr. 2.903; Hor. Sat. 1.8.17) was rejected and ad- was 
interpolated so that the line scanned ārma ādsuētă. This seems less likely to me. Likewise, there is no 
need to add -que, as Heinsius suggests. 



 
 

Sc. Emathii. Heinsius probably thought that the transmitted text was surprisingly 
unspecific and missed a reference to one of the rivers in Thessaly, where Chiron lived.46 
But Ocyr(rh)oë was named, so to speak, after the flumen rapidum (ὠκύ- + ῥοή) in which 
she was born. 
 
2.677 nam nec] nec enim (69v)47 
 
Heinsius (1659) 54 and 43n edited sed nec with four mss.,48 but he also conjectured nec 
enim, which is the reading of G,49 a ms. he did not know. It is not easy to reach a 
decision, but nam nec is probably right. 
 
2.687 notus] fotus (69v) 
 
Notus is perhaps a bit pleonastic, considering the parenthesis in 688. But fotus is no big 
improvement.50 
 
2.693 facto] pacto (70r) 
 
The confusion between factum and pactum is documented, for instance, at 11.135 or 
Fast. 4.817. Facto could certainly be a gloss for pacto. Heinsius referred to 6.429 (non 
Hymenaeus adest, non illi Gratia lecto), where he also proposed pacto (b3, 175v), and 
to 7.700 sacra iugalia, where he read pacta (1659, 163). However, the possible 
‘agreement’ is actually not suggested until the next line. 
 
2.709 gratamque] sanctamque (70v) 
 
Heinsius made his suggestion after anctamq. (sic) of “pr. Gron.” [B4ac, an auctamque?]. 
But I have been unable to find examples in Ovid of sanctus applied to places other than 
temples. For gratus, cf. 2.758 and see Bömer (1969) 405‒406. 
 
2.720 agilis] auidos (71r) 
 
Slater (1927) ad loc. already documented this conjecture. I add that in b3 Heinsius 
himself remarked: “male. nam avidus praecessit [719]”. 
 
2.744 uerba] certa (71r) 
 
Lr26 has uisa per auras in 743, whence Heinsius suggested certa for uerba here. The 
change would ironically point out that Mercury is not actually acting as Jupiter’s envoy 
but following his own desire for Herse. But cf. Am. 2.5.20 uerbaque pro certis iussa 
ualere notis; Fast. 2.509‒510 iussit et in tenues oculis euanuit auras; / conuocat hic 
populos iussaque uerba refert. 

 
46 Similarly, Heinsius (1659) 42n attempted patrii, making thus Chariclo daughter of the river. 
47 Also o4, 51. 
48 I have been unable to identify them. 
49 Bo3 has neque enim, but, as far as I can tell, Heinsius did not take notice of the reading in o4, where his 
collation is extant. 
50 What are we to make of totus of Li3 and solus of Go? Could they point to the palaeographically close 
fotus? 



 
 

 
2.755 hanc] hinc (71v) 
 
But hanc is probably necessary, just to make clear that the subject of detexisse is 
Aglauros. 
 
2.789 totum] s(a)eptum (73r) 
 
Considering readings such as comptum (“unus Med.” [Lr22]), circum (“Junian.” [Et]), 
rectum (“Sixi” [B7, f. 20v (an tectum?)]), Heinsius suggested s(a)eptum, which is 
possible (vid. OLD s.v. 6) and enlivens the description of Invidia’s extraordinary attire. 
Cf., although it is a different example, Ars am. 3.662 cingenda est altis saepibus ista 
seges. But for totum, cf. Fast. 5.335 tempora sutilibus cinguntur tota coronis (where, 
nonetheless, tota has raised some doubts); cf. also Met. 5.388 cingens latus omne. 
 
2.808 incerto] infesto (74r) 
 
Heinsius attempted infesto, which does not go too well with lenta ... tabe, but is still 
better than Merkel’s (1875) IX inserto (coll. Lucr. 2.115). For incerto, see Bömer 
(1969) 427. 
 
3.75 quique] quaque (80r) 
 
Heinsius’s quaque, inspired by the reading quoque (vid. Suárez 2015, 523), is possible 
(cf., e.g., 1.241; 5.290 “qua”que “uia est uobis, erit et mihi” dixit “eadem”), but 
unnecessary (cf., e.g., 1.133‒134 quaeque diu steterant in montibus altis / fluctibus 
ignotis exsultauere carinae). 
 
3.84 figitque] triuitque (80r) 
 
Heinsius edited frangitque (vid. Suárez 2015, 82‒83), but he had also suggested 
triuitque adducing 8.369 dentibus ille ferox [sc. aper] in querno stipite tritis (“ubi pro 
dentibus tritis forte fractis legendum”). It is an interesting halfway between figit and 
frangit. 
 
3.97 quid] quod (81r) 
 
Quod should introduce a causal clause dependent of et tu spectabere serpens. But this 
seems inferior to the vulgate. 
 
3.190 perfudit] perfundens (84r)51 
 
Heinsius’s perfundens, apparently inspired by perfundit (“tres libri [V8 Lr9ac] ... cum pr. 
Regio [P4] & pr. Cantabr. [deperditus]”),52 could be right at first sight. But the series of 
enclitics of 189‒190 coordinates finite verbs. 
 
3.191 cladis] cladi (83v) 
 

 
51 Also o4, 64. 
52 Vid. n. 11. 



 
 

A dative could go well with addidit, while the genitive with praenuntius (vid. Bömer 
1969, 500) is expendable (cf. 15.670). But Diana’s words do indeed announce the 
clades; we must retain the genitive.53 
 
3.206 primique] primoque (84r) 
 
Heinsius suggested primoque as an alternative to primumque,54 which he had found in 
M. But the predicative is clearly to be preferred.55 
 
3.242 hortatibus] ululatibus (84v) 
 
Heinsius adduced 15.797 nocturnos ululasse canes in support of his proposal, which is a 
different example. Here, Actaeon’s former companions are urging the dogs. Cf., e.g., 
10.537; Her. 4.42. 
 
3.247 uidere] tueri (85r) 
 
Heinsius knew the erroneous reading uideri of some important mss. and cleverly 
conjectured tueri: “quae vox saepe corrupta apud nostrum et mutata in uideri lib. I. 85. 
ex Mss. [read Burman 1727, 21]”. Could uidere be an emendation of uideri, which in 
turn originated in tueri?56 
 
3.281 ut sit] absit (86v) 
 
Heinsius’s absit, after adsit of “duo libri” (Lr21 Pt3?),57 is not impossible. But Juno, 
disguised as Semele’s wet nurse, encourages her to ask Jupiter for prove of his identity 
(cf. 281‒283). Cf. also Hyg. Fab. 167.2. 
 
3.355 tetigere] pepigere (90r) 
 
An allusion through pepigere to marital engagements would be enticing, but Narcissus 
can hardly be the object of the verb. 
 
3.368 haec] hac (90r) 
 
This derives from hoc of P41. But hac in fine does not make sense. Haec is certainly 
required to signal the change of subject. 
 
3.384 recepit] receptis (89v) 
 
After recepto of “unus Leid.” (Ld71) Heinsius thought of receptis, which precludes 
punctuating before perstat. This is appealing: the cause of Narcissus’ astonishment in 
385 is none other than his own repeated words. Verba, which must now be construed 

 
53 See also n. 34. 
54 For primum and primo, vid. Bömer (1976) 376 on 5.592. 
55 I do not fully share Suárez’s (2015) 181 views. I think scribes must have assimilated primi to the 
closest noun or replaced it with an adverb. Cf. Fàbregas (2022) 233 on 10.240. 
56 For tueri facta, cf., e.g., Verg. Aen. 10.397‒398. But see Suárez (2015) 214 for another, more 
straightforward, probably better, interpretation, of how uideri might have originated. 
57 In o4, 67, Heinsius assigns the reading to g (= Pt3). I believe that a folium, between ff. 31 and 32, once 
containing 3.232‒295 is now missing; Heinsius certainly still saw it. 



 
 

with quot dixit, could have been misunderstood and the participle altered into a personal 
form.58 Cf. on 3.645. 
 
3.410 turbarat] turbarant (92r) 
 
Turbarant actually appears in a couple of mss. unknown to Heinsius (vid. Suárez 2015, 
591). Both forms are possible, but the postverbal position of one of the subjects and the 
emphatic coordination call for the singular, as transmitted (cf. Pinkster 2015, 152‒153 
[§13.6.v]). 
 
3.456 est mea quam] sunt mihi quas (93v) 
 
Again, Heinsius wanted an agreement in the plural. But see on 3.410; also Suárez 
(2015) 332. On the other hand, the dative is attractive (cf., e.g., 1.192‒195; 13.286; Her. 
3.80; 3.146; 5.86; Ars am. 3.205‒206; Fast. 6.219‒220).59 
 
3.458 ultro] et tu (94r) 
 
Cf., e.g., Hor. Sat. 2.3.322 quae si quis sanus fecit, sanus facis et tu. Ovid has et tu at 
the end of the line at Her. 10.5. However, we are talking about Narcissus’ image, so 
ultro (vid. OLD s.v. 5 “of one’s own accord”) is certainly more pointed. 
 
3.459 saepe] et ipse (94r) 
 
This proposal, which tries to adjust the unmetrical ipse of Lr instead of saepe, is really 
not bad (cf., e.g., Verg. Aen. 1.5 multa quoque et bello passus). 
 
3.477 quo refugis? remane] quo fugis? ah remane (93v) 
 
As an alternative to his metrically objectionable quo fugis? oro, mane (see Suárez 2015, 
342) and to the attractive quo fugis? o remane of Ld6,60 Heinsius had also envisaged 
quo fugis? ah remane,61 which is very good. Cf. Verg. Ecl. 2.60 quem fugis, a, 
demens?; Prop. 2.30.1 quo fugis a demens?; cf. also Ov. Am. 1.13.3 quo properas, 
Aurora? mane! Ovid similarly combines fugere and remanere at Pont. 1.6.29‒30 haec 
dea [sc. Spes], cum fugerent sceleratas numina terras, / in dis inuisa sola remansit 
humo.62 
 
3.478 quod] quem (93v) 
 
Heinsius wrote: “malo &, quem”. Cf., e.g., 10.387 (with Fàbregas 2022, 292). But the 
transmitted quod is probably right: cf. Am. 1.4.58; Ars am. 1.91‒92; Her. 18.179. 

 
58 A delayed antecedent is certainly expected. On the one hand, this might explain the change in the 
transmission; on the other, it discredits the proposal. 
59 In his first edition Heinsius (1652) 58 printed est mihi, quam. See also n. 34. 
60 Cf. Verg. Aen. 10.369 quo fugitis, socii?; 10.649; Ov. Her. 7.41; 10.35; Rem. am. 580; Met. 8.108; 
8.110 quo fugis immitis ...? (note crudelis at 3.477); cf. also Ars am. 1.124 haec manet, illa fugit. It is not 
unlikely that o fell out and the gap was filled back in with re-fugis, being re- readily available in re-mane. 
61 One might wish to write a. See Ramírez de Verger (2021) 202 on 6.386; Fàbregas (2022) 383. 
62 As for the transmitted text, Bach (1831) 144 objects to the “Missklang” re- re-. I would like to see a 
similar example of adjacent verbs with the same prefix and opposed meaning. I have not found good 
parallels for the combination of refugere and remanere either. 



 
 

 
3.480 dumque dolet] dumque dolens uel dum dolet et (95r) 
 
Both proposals are palaeographically possible and extend the dum clause up to the end 
of line 481. But the text as transmitted is clearly more fluent.63 
 
3.518 ne] neu (95v) 
 
No doubt, the sense would admit a disjunctive within the si-clause.64 But I have serious 
doubts that neu(e) is grammatically correct in this context.65 
 
3.523 tuam] tuo (96r) 
 
This is a significant improvement. It is of no great consequence whose blood defiles the 
woods, while it is all but trivial to insist that Pentheus’ blood will sully his own aunts 
and mother. Of course, the delayed possessive might have been easily assimilated to 
matrem. For a similar hyperbaton (if indeed hyperbaton it is), cf. 5.281‒282 “nec 
dubitate, precor, tecto graue sidus et imbrem” / (imber erat) “uitare meo. See also Her. 
4.58; Am. 2.4.36; Her. 16.215‒216; Ib. 360. Additionally, cf. Octavia 148‒149 Silanus 
et cruore foedauit suo / patrios penates. 
 
3.538 qui] quos uel quam (96v) 
 
Heinsius commented: “Si miseri legamus [pro mirer], cum Lactantio et aliis [cf. Suárez 
2015, 377], etiam quos longa vel quam erit legendum, ut ad Tyron aut ad penates 
referatur”. This is not bad, even if we retain mirer. The relative might have been easily 
attracted to qui, being placed between senes and uecti (and miseri). 
 
3.539 hac Tyron, hac profugos] a Tyria hac profugos uel hac Tyrios profugi (97r) 
 
Taking advantage of the reading a Tyrio hac of “duo alii” (H3 Lr9)66 and a Tyrio ac of 
“alius” (Lr19), Heinsius attempted a Tyria [sc. sede] hac [sc. sede].67 But he right away 
stated, correctly: “Sed Hac Tyron, hac vera lectio est”. However, he also came up with 
another proposal (“posset legi”): Hac Tyrios profugi, which is interesting. Here the 
senes, not the Penates, are profugi, just like Aeneas (vid. Bömer 1969, 579). But the 
double epithet uecti/profugi is not very elegant.68 Furthermore, cf. Prop. 4.1.39 huc 
melius profugos misisti, Troia, Penatis (with Fedeli/Dimundo/Ciccarelli 2015, 229). 
 
3.554 iuuant] iuuent (96v) 
 

 
63 One should also consider that the perfect is possible in dum clauses, but it often implies a “contrast 
between the events in the main and subordinate clauses” (Pinkster 2015, 615) which seems to be absent 
here. 
64 This would be better than the available options (i.e. a final/consecutive clause with ne or a negative 
copulative clause with nec). 
65 Neu(e) is mostly used in jussive contexts and is extremely rare with the imperfect subjunctive (vid. 
Lease 1913, 418; Szantyr 1972, 331). 
66 Another one was probably H (vid. Suárez 2015, 617). 
67 This proposal also features in o4, 76. For the alteration a/hac, cf. Rivero (2018) 53‒54 on 13.8. 
68 The repetition hac ... hac seems also more expressive. Cf. Wills (1996) 76‒79. 



 
 

Although the subjunctive in the relative clause might further characterize Pentheus’ 
contempt for the puer inermis, there is no real need for the change. 
 
3.585 non ulla armenta] armentaue multa (97v)69 
 
Heinsius’s proposal derives from armentaque nulla of “unus Basil.” (Bs3; also in some 
other mss.; see Suárez 2015, 405). One could perhaps approve of -ue (corresponding to 
lanigerosue),70 but multa suggests that Acoetes inherited at least some livestock, which 
is impossible for a man tam pauper ut cubili careat.71 Cf. 3.590‒591. 
 
3.595 notaui] notari (99r) 
 
Heinsius anticipated Watt’s (1995) 98 very alluring notare. The passive infinitive72 
seems unnatural after flectere (whose subject is inferred from addidici), but its intrinsic 
difficulty could more easily explain why it was substituted for a palaeographically close 
notaui. Suárez (2015) 408 tried to defend the transmitted notaui arguing, 
unconvincingly, that sidera notare requires less skill than regimen flectere. 
 
3.600 nox ibi] noxque ibi (99r) 
 
Heinsius (1659) 76 kept the vulgate nox ubi, but he knew the alternative nox ibi, whence 
he proposed noxque ibi, continuing the double -que of 599 (the phrase is certainly 
linked to it). The subsequent Aurora rubescere primo / coeperat initiates a new period 
and a new stage of the tale. For the hypothetic loss of -que in synaloepha, vid. Fàbregas 
(2022) 112. 
 
3.624 de] e (100r) 
 
This is not impossible (cf. 4.791; Ars am. 3.447), and is actually the reading of P5. But 
cf. 9.70‒71 nec ullum / de comitum numero caput est impune recisum; Fast. 1.548; Tr. 
5.3.44. See also on 4.516. 
 
3.636 “Naxon” ait Liber] “Naxon” Liber ait (99v) 
 
Not a completely negligible suggestion, since Ovid usually places the subject of ait 
before the verb: cf., e.g., 3.10 Phoebus ait (6.215; 10.197); 3.532 Pentheus ait; 4.572 
Cadmus ait; 12.201 Caenis ait; Fast. 2.759 coniunx ait; 2.796 natus ait regis; 6.444 
Metellus ait. However, for the transmitted order: cf. Met. 4.639 ait Perseus illi; 8.350 
ait Ampycides; 12.490 ait Caeneus; Rem. am. 475 ait Atrides. 
 
3.644 “capiat”que “aliquis] “capias”que “alias (100r) 
 

 
69 Also o4, 77. 
70 However, for the asyndeton, cf. 7.558 and Verg. Aen. 8.376 (adduced by Suárez 2015, 405); perhaps 
also Met. 1.512‒514 (quoted by Bömer 1969, 591). 
71 Burman (1727) 222 on 3.582 uses more or less these terms. Fishermen were proverbially very poor (cf., 
e.g., Verg. Aen. 12.517‒520). 
72 Cf., e.g., Pont. 2.10.22 Trinacris est oculis te duce nota [uisa v.l.] meis. 



 
 

Heinsius’s proposal probably tries to make sense of capiasque aliquis of the “Lovan.” 
(capiatque aͥs Br2 (alius Br22s), f. 42v; also B), but it is absurd for Acoetes to tell 
Opheltes that he may take the rudder alias and yield it in the very next line. 
 
3.645 remoui] remoto (99v) 
 
Heinsius attempted remoto. This would more clearly state why Acoetes’ companions 
“cried out upon him” (we should remove any punctuation after the participle). Cf. on 
3.384. 
 
3.650 tum deus inludens] tum deus, ille, deus (99v) 
 
The proposal is palaeographically great (cf. Lucr. 5.8 deus ille fuit, deus; Verg. Ecl. 
5.64; Stat. Silv. 4.6.36), but I see no reason to suspect i(n/l)ludens (cf., e.g., 652 flenti 
similis; cf. also 9.66 risit et inludens nostras Tirynthius artes). 
 
3.672 et expresso] de impresso (100v) 
 
Adopting the reading pinnis in the previous line, and probably combining readings such 
as depresso and impresso, Heinsius attempted de impresso, which is paleographically 
apt (and certainly better than a two-fold bare ablative). But, to avoid describing the 
same process twice (corpore de impresso / curuamine flecti), et is necessary (Suárez 
2015, 457) and the participle must agree with curuamine.73 
 
3.708 monte fere medio est] monte viret (medium (101v) 
 
After medium est of “Thuan. opt.” (P28), Heinsius suggested monte uiret (medium 
cingentibus ultima siluis), which, despite being palaeographically clever, looks like an 
unnecessary, syntactically ambiguous alteration.74 See also on 3.709. 
 
3.709 purus ab arboribus] rura, sine arboribus, (101v) 
 
Perhaps Heinsius disliked the presence of two epithets with campus, as well as purus 
with a(b), which is rare (Bömer 1969, 621).75 Thus, he tried to introduce a noun that 
could be construed with ultima (on which see Bömer ibid.). But I believe that rura does 
not fit well into the description of the campus. See also on 3.708. 
 
3.713 et ‘o geminae’] ‘io Ogygiae’ (100v) 
 
Considering the reading io of some codices, Heinsius thought of io Ogygiae, as an 
alternative to his et Ogygiae (1659, 75n). This is very good, but one would rather expect 
a hiatus, as in 5.625. 

 
73 In that case, if Oellacher is right in his interpretation (ThlL V 1785.34‒43 “ex corpore extrudere, 
emittere”), expresso suggests a sort of expansion of the spine, while perhaps impresso hints at a forced 
arching (cf. Stat. Theb. 5.403‒409 Iason ... / transtra per et remos impressaque terga uirorum / ... / uoce 
manuque rogat). However, I would not rule out a technical or artistic use of exprimere here (vid., e.g., 
OLD s.v. 6). 
74 For fere, see Bömer (1969) 620‒621. 
75 A similar textual problem is to be found at 13.910‒911 (see Rivero 2018, 387‒389). 



 
 

 
3.716 fremituque] unumque (101v) 
 
Heinsius independently suggested unumque (at least, he makes no reference whatsoever 
to Capoferreus 1659, ad loc.). See Suárez (2015) 485. 
 
3.724 ostendens] obtendens (101r) 
 
The compound would pick up and pathetically intensify the previous tendat. See Wills 
(1996) 443‒445. The confusion between ostendo and obtendo is, for obvious reasons, 
not uncommon (ThlL IX.2 273.68). However, the verb usually entails the notion of 
covering or protecting, which seems out of the question here. On the other hand, 
ostendens matches aspice and uisis of 725. 
 
3.732 exemplis monitae] extemplo motae (103r) 
 
Burman (1727) 234 claims credit for this proposal, but it should be attributed to 
Heinsius. 
 
4.61 patres] parens (105v) 
 
After pares of “m. pr. Urbin.” (V21), Heinsius attempted parens (coll. 4.155), which 
looks wrong with uetuere. For patres, see Bömer (1976) 39‒40. 
 
4.61 quod ... uetare] quae ... uetari (uel quas ... uetare) (105v) 
 
The former was later suggested by Bothe (1818) 32. Quae, referred to taedae, was 
probably misunderstood and substituted for quod. Or, as Bothe implies, the infinitive 
could have been assimilated to uetuere and potuere, and the syntax adjusted 
accordingly. Cf. 12.546 quod mallem posse negare (Tarrant 2004, 366: “negari hp2 
(Heinsius)”). Heinsius took inspiration from qui of “alius Medic.” (Lr121). His second 
alternative, less attractive, though involving a slighter change, derives from quos of 
“sec. Medic.” (Lr8). 
 
4.150 est et ... hic] huic et ... haec (107v) 
 
Apparently, Heinsius drew inspiration from hoc et ... haec (“pr. Vatic. [V5, f. 33v, reads 
est et ... haec] & Thysii [unidentified]”), hoc et ... hoc (“unus meus” [O42v]), and 
perhaps huic et ... hic of Lr12 (not quoted in b3). Huic, referring to manus, is not bad, 
but what are haec uulnera? Pyramus has just said that Thisbe was lost by her manus 
amorque; in turn he must rhetorically proclaim to have both as well. For the repetition 
of est et, see Bömer (1976) 61. 
 
4.207 ipsoque] ipsasque (110v) 
 
“Locus mendi omnino suspectus” (Heinsius 1659, 83n), in part maybe because of the 
multiple variant readings. The expression ipso illo tempore is prosaic and “pedantisch”, 
as Bömer (1976) 82 calls it. Heinsius suggestion “per appositionem”, with great 



 
 

palaeographical economy, clearly establishes Clytie’s rivals as the cause of her graue 
uulnus and of her grief,76 which will eventually lead to her transformation (4.256‒273). 
 
4.261 nudis] passis (112v)77 
 
After the reading sparsis of some mss. (“IV nostri [Ld2 O12(teste Heinsio)78Lr23]”), 
Heinsius put forward, again, passis. See on 1.542. But Heinsius himself compared 7.183 
nuda pedem, nudos umeris infusa capillos. Cf. also Her. 14.100 diues eras, nuda nuda 
recumbis humo. In this case, it rather seems that nudis was lost after nuda and the gap 
filled with sparsis (which is a bit pleonastic with incompta). 
 
4.269 tenetur] retenta (113r) 
 
Retenta, which implies the notion of stopping or refraining, seems to suggest that 
Clytie, now sunflower, would follow Helios, if she could. Cf., e.g., 2.348‒349 ad quam 
conata uenire / candida Lampetie subita radice retenta est. For quamuis with a 
participle or adjective, cf. 4.206; 3.494; 4.232; 9.476. But for teneo, cf. 11.77‒78 at 
illam / lenta tenet radix. 
 
4.287 fontis] fonti (115r) 
 
The possessive dative is possible and could have been replaced by the genitive, a lectio 
facilior.79 But, in truth, there is nothing wrong with the transmitted text (cf., e.g., 
15.299; Tr. 3.10.17 tantaque commoti uis est Aquilonis ut ...). 
 
4.288 diua Cythereide] diuaque Cytheride (114v) 
 
The attempt to correct diuaque Cythareide of “unus Leid.” (diua(que) cithareia Ld8, f. 
35r)80 is futile. Mercurio is a dative, not an ablative (as whoever was responsible for the 
addition of -que surely thought): cf., e.g., Caes. BGall. 5.14.5; Mart. 6.39.13. The form 
Cythereis recurs at Manil. 2.33.81 
 
4.328 seu] fac (115v) 
 
It seems that Heinsius wanted to avoid the repetition of seu (321; siue 322 and 327). 
Fac should mean “suppose” (OLD s.v. 20), but this requires a subjunctive or an 
infinitive. Should we understand fac nullam esse? 
 
4.341 ut puer et uacuis ut] ilicet ut uacuus et (116r) 
 
Heinsius suggested this as an alternative (“vel”) to his excellent ut puer et uacuis ut, 
probably trying to stay closer to the text of most mss. (scilicet ut uacuus et). 

 
76 “Vulnera sunt quoque amantium perfidiae et riualium simultates” (Pichon 1902, 302). 
77 Also o4, 91. 
78 o4, 91. 3.397‒4.358 were lost after Heinsius collated O12 (Reeve 1974, 153 n. 78). 
79 See also n. 34. 
80 This is not Ld2 (f. 37v) nor Ld7 (f. 36r). 
81 See Bömer (1976) 109. Cf. also Pelopeis (Fast. 4.285) and Pelopeias (Fast. 3.83). The form Cytheris 
lacks authority referring to Venus (ThlL O.2 811.75‒812.28). 



 
 

 
4.359 tangit] plangit (117r)82 
 
The confusion between the two verbs is frequent (e.g. 2.584; 3.125 [with Suárez 2015, 
116‒117]; 11.81 [with Díez 2014, 62]), and pectora plangere is likewise a common 
expression (ThlL X.1 2310.66‒2311.6), but it is usually a display of grief. For tangere, 
cf. 8.605‒606 ipse natantis / pectora tangebam trepido salientia motu; Fast. 2.803‒804 
positis urgentur pectora palmis, / tum primum externa pectora tacta manu; 6.753. 
 
4.371 et] ita (116v) 
 
The proposal was probably triggered by ut of P4 O4 P24. The repetition of ita is not 
unattractive. See ThlL VII.2 527.20‒29. But apparently Ovid only admits synaloepha of 
ita over est.83 
 
4.373 nam] iam (116v) 
 
The confusion is common enough (see Suárez 2015, 1 on 4.1; Fàbregas 2022, 431), but 
nam is not wrong here: cf. 10.489. 
 
4.375 conducat] conductos (117r) 
 
Heinsius (1659) 91 and 89n opted for conducta, but he had also suggested conductos, 
which seems also possible and perhaps even stylistically superior. The finite verb could 
have been easily introduced after si quis. But the change is not truly necessary. See 
Bömer (1976) 129. 
 
4.463 undas] urnas (120v) 
 
Heinsius’s proposal seems to derive from umbras of O42v. But quas perdant guarantees 
undas: cf. Ib. 177 quaeque gerunt umeris perituras Belides undas. 
 
4.499 mens est, quae diros sentiat ictus] mens quae letalem sentiat ictum uel mens est, 
quam diro sauciat ictu (122r) 
 
Heinsius based his first proposal on tales instead of diros of Lr20. For letalis ictus, see 
ThlL VII.1 169.12. The conjecture is unnecessary: cf. Her. 21.135 quae iurat, mens est. 
The second conjecture derives from sauciat of “Sixi” (B7), but one would rather expect 
sauciant, which is unmetrical. 
 
4.516 deque] eque (121v) 
 
On the usage of de, see Loewe (1889) 6‒7. There is no need for the conjecture. See also 
on 3.624. 
 
4.520 fecit] faciat (121v) 
 

 
82 See also Rivero (2018) 410‒411 on 13.965, where Heinsius suggested frangeris instead of tangeris. 
83 Pont. 1.1.37; 4.4.3. But cf., e.g., ita illum (Plaut. Men. 41); ita istaec (Plaut. Men. 94); ita ad (Verg. 
Aen. 10.459); ita ut (Hor. Sat. 1.2.20). 



 
 

Perhaps Heinsius wanted to avoid uariatio temporum, but Ovid, as is his wont, presents 
two possible causes of why Ino is concita. The indicative is grammatically correct with 
seu. 
 
4.530 recanduit] ecanuit (123r) 
 
Recandesco (?) is an ‘Ovidian’ verb, which he uses in the sense of “to become white” 
only here and, perhaps, at 7.7884 (later Claud. 15.211). Heinsius coined a different 
compound, ecanuit, for this line.85 E- could have been lost in synaloepha, replaced with 
re- and -d- added (percussa͜͜‿ēcānuit > percussă cānuit > percussă rĕcanduit). A pity 
that the verb is not otherwise attested!87 
 
4.620 infestaque] infestis (126r) 
 
Heinsius seems to have disliked for a moment the double epithet with terra. The 
‘active’ sense of infestus is certainly possible. But, on second thought, Heinsius himself 
crossed the proposal out. See Bömer (1976) 191.  
 
4.654 tibi gratia nostra est] uictoria constat (127v) 
 
Heinsius’s drastic emendation has at least some diagnostic value. Perseus’ utterance is 
certainly not as sharp as in 5.225: why on earth would he give a ‘boon’, even an ironical 
one, to Atlas for denying him the hospitality his gratia (OLD s.v. 5) as a hero deserves? 
However, reading uictoria constat, Atlas would receive a ‘prize’ for winning easily, and 
the sarcasm of Perseus’ utterance would be stronger. However, I cannot conceive a 
plausible iter corruptionis. 
 
4.657 mons factus] mutatus (127v) 
 
This proposal derives from mons tantus of P2 and “unus meus” (quantus Atlas o mons 
tantus erat O12, f. 41v).88 Mons factus could be a gloss of mutatus, but I still see no 
cause for the emendation. Tantus certainly derives from quantus. 
 
4.701 cunctis certe] numquid cunctis (129v) 
 
Heinsius suggested praeferrer numquid cunctis gener?, reporting that Lr8 (f. 54v) reads 
the text of most mss. “cum interrogationis nota”. The interrogative is attractive, but the 
negative answer usually implied by numquid is not adequate. 
 
4.718 celeri missus praeceps] celer admisso praepes (131r) 
 
Magnus (1914) 160 attested Heinsius’s suggestion celeri admissum praeceps in b3,89 
but he failed to record this additional (“posset et legi”) and better option, which, beyond 

 
84 Holmes (ThlL XI.2 264.1) takes it so, but this is only a possibility; see Kenney’s note ad loc. (Barchiesi 
2024, II, 44). 
85 For a similar compound, cf. erubesco, excandesco. For other Ovidian verbal compounds in e-, see 
Draeger (1888) 18 and Linse (1891) 53. 
87 One could perhaps think of incanuit, which is at least attested at Prisc. Inst. 2.438.5. However, both 
options sacrifice the anagram recanduit unda; see Rosati ad loc. (Barchiesi 2024, I, 429). 
88 O42m has tantus, but it is unclear what it refers to; P93s has tantus over mons. 



 
 

the alternation praepes/praeceps,90 anticipates Slater’s (1927) ad loc. celer admisso. 
Slater compared 1.532 admisso sequitur uestigia passu. See Bömer (1969) 166. 
 
4.757 tanti] sancti (132r) 
 
Andromeda’s rescue was part of an agreement between her father and Perseus. In this 
sense, sancti is possible.91 But there is nothing wrong with tanti ... facti: cf. Verg. Aen. 
7.232 tantique abolescet gratia facti; Ov. Tr. 5.4.47 tantis referetur gratia factis. 
 
4.761 tibiaque et cantus] tibiaque cantusque (132r) 
 
An interesting proposal, continuing the series of -que initiated in the previous line with 
lyraeque. Cf. 3.529‒530; 4.9‒10; 8.526‒527. But perhaps cantus introduces a new 
category. See Fàbregas (2020); vid. also on 4.778. 
 
4.767a unus] orsus (131v) 
 
The vulgate Heinsius received of the troubled lines 767‒769 was: 
 

cultusque genusque locorum 767 
quaerit Abantiades; quaerenti protinus unus 767a 
narrat Lyncides, moresque, animosque uirorum. 768 
quae simul edocuit ... 769 

 
After ursus of four mss. (e.g. Lr14, unus Lr14v), Heinsius attempted orsus, “si vulgata 
teneatur”. This removes the objectionable unus,92 but, if both Abantiades and Lyncides 
(cf. 5.99) allude to Perseus, the passage is still absurd. The solutions published by 
Heinsius (1659) 106‒107n, involving some kind of deletion, are certainly better.93 
 
4.770 quibusque] quibusue (132r) 
 
Quibusue in a similar context and at the end of the line appears at Verg. Aen. 6.560. Cf. 
also Ov. Tr. 3.12.34. For the scribal tendency to substitute -ue for -que, see Rivero 
(2018) 414; Fàbregas (2022) 204‒205. 
 
4.771 artibus] astibus (132r) 
 
This is a very clever proposal. Of course, a rare term such as astibus could have been 
easily replaced by artibus, an obvious palaeographic rival. However, this dative-ablative 
plural does not seem to be attested anywhere in classical literature, although the singular 
astu was certainly in use. In fact, it appears in line 776. For artibus, cf., e.g., 9.62. 

 
89 Magnus does not say Heinsius adduced Verg. Aen. 11.272 At socii admissis [instead of amissi] 
petierunt aethera pennis (“sic enim cum vetustissimo libro legendum”, by which he does not refer to any 
of the late-antique mss.; cf. Heinsius ap. Burman 1746, III, 630). 
90 About it see Heinsius’s note and Burman’s (1727) 304‒305 objections. 
91 As reported by Burman (1727) 308, Heinsius also suggested tanti ... pacti. See also on 2.693. 
92 The use of unus as quidam has some informal or late Latin aroma, even if Bach (1831) 210 compares 
Caes. BGall. 5.45.2 erat unus intus Neruius nomine Vertico. See also Oudendorp (1737) 266 ad loc. Read 
Szantyr (1972) 193. In Ovid, cf. Her. 15.162. 
93 Lines 767a and 768, missing in some mss., are now regarded as spurious, but I am not quite convinced 
this settles all the problems of the sequence. See also Anderson (1997) 493. 



 
 

 
4.778 deuiaque et siluis] deuiaque siluisque (132v) 
 
This is much the same proposal as on 4.761. But here one would rather prefer that the 
double enclitic coordinated two exactly equivalent words. 
 
4.786 matris de] materno (132v) 
 
This is the reading of Dr. Cf. 5.259 uidi ipsum [sc. Pegasum] materno sanguine nasci. 
Cf. also Lucr. 4.1214; Tib. 1.2.41. See ThlL IX.1 82.33‒54. But maybe the uariatio of 
construction as regards 5.259 is correct. For nasci with de, cf., e.g., 2.469; 15.836. See 
ThlL IX.1 82.7‒23. 
 
4.796‒797 illa, neque in tota conspectior ulla capillis / pars fuit] illa nec tortis 
conspectior ulla capillis / parue erat (uel claruit) uel illi nec toto conspectis orbe capillis 
/ par fuit (133r) 
 
Introducing the adjective par instead of pars is certainly attractive (cf. Her. 16.143‒144 
nec tibi par usquam Phrygia nec solis ab ortu / *** / inter formosas altera nomen 
habet; Fast. 1.90).94 For the postponement of nec, see Platnauer (1951) 93‒96. The 
second conjecture tries to make sense, through extensive rewriting, of the reading urbe 
capillis of “unus Medic. ex melioribus” (Lr7) and toto of “duo” (Lr20O12ac(ut uid.)). 
The former is perhaps more subtle, avoiding so many ablatives, but I must object to 
tortis ... capillis, because the transformation of Medea’s hair has not taken place yet.95 
Reading totis with some important mss. would be no great improvement.96 
 
Acknowledgement: I am grateful to Álvaro Cancela, Antonio Ramírez de Verger and 
Luis Rivero for having commented on a draft of this paper, and to Juan Antonio Estévez 
and Jaume Juan for valuable advice on specific points. Thanks are also due to the 
editors of Philologus and to the anonymous referees for saving me from some errors. 
This paper is part of the research project PID2019-106844GB-I00 financed by the 
Spanish Ministry of Science and Innovation. 
 
Bibliography 
 
Anthologia veterum latinorum epigrammatum et poematum ... cura Petri Burmanni Secundi, I, 

Amstelaedami 1759. 
C. Julii Caesaris De Bellis Gallico et Civili Pompejano, nec non A. Hirtii, aliorumque De Bellis 

Alexandrino, Africano, et Hispaniensi Commentarii ... Cura et studio F. Oudendorpii, 
Lugduni Batavorum et Rotterodami 1737. 

M. Manilii Astronomicon liber quartus, recensuit et enarrauit A. E. Housman, Londini 1920. 
P. Ovidii Nasonis Operum tomus II qui XV. Metamorphoseon, sive Transformationum, libros 

continet, Amstelodami 1652. 
P. Ovidii Nasonis Operum tomus I scripta amatoria complexus, Nicolaus Heinsius, Dan. Fil. 

infinitis locis castigavit ad fidem scriptorum exemplarium, Amstelaedami 1658. 
P. Ovidii Nasonis Operum tomus II qui Metamorphoses complectitur, Nicolaus Heinsius D. F. 

locis infinitis ex fide scriptorum exemplarium castigavit et observationes adiecit, 
Amstelaedami 1659. 

 
94 On the confusion, see ThlL X.1 261.36‒37. 
95 Cf. Hor. Carm. 2.13.35‒36 [quid mirum, ubi] intorti capillis / Eumenidum recreantur angues? 
96 On the other hand, the repetition clarissima (794)/claruit is not enticing. 



 
 

P. Ovidii Nasonis Metamorphoseon libri xv cum integris ... notis, quibus et suas adnotationes 
adiecit Petrus Burmannus, Amstelodami 1727. 

P. Ovidii Nasonis Metamorphoseon libri XV, mit kritischen und erläuternden Anmerkungen von 
E. C. Chr. Bach, Erster Band I‒VII, Hannover 1831. 

P. Ovidius Naso Metamorphoses ex iterata R. Merkelii recognitione cum emendationis 
summario, Lipsiae 21875. 

P. Ovidii Nasonis Metamorphoseon libri XV, Lactanti Placidi qui dicitur narrationes fabularum 
Ovidianarum, ed. H. Magnus, Berolini 1914. 

P. Ovidi Nasonis Metamorphoses, rec. R. J. Tarrant, Oxonii 2004. 
P. Virgilii Maronis Opera, cum integris et emendatioribus commentariis Servii, Philargyrii, 

Pierii. Accedunt ... praecipue Nic. Heinsii notae nunc primae editae; quibus et suas in 
omne opus animadversiones et variantes in Servium lectiones addidit P. Burmannus, 
post cuius obitum interruptam editionis curam suscepit et adornavit P. Burmannus 
junior, Amstelaedami 1746. 

 
W. S. Anderson, Ovid’s Metamorphoses Books 1‒5, Norman 1997. 
A. Barchiesi (ed.), A Commentary on Ovid’s Metamorphoses, 3 Vols., Cambridge 2024 (Milano 

12005‒2015). 
E. Bednara, “De sermone dactylicorum Latinorum quaestiones”, Archiv für lateinische 

Lexikographie und Grammatik 14, 1906, 317‒360; 532‒604. 
F. F. Blok, Nicolaas Heinsius in dienst van Christina von Zweden, Delft 1949. 
H. Boese, “Zu den Ovidkollationen des N. Heinsius”, Philologus 106, 1962, 155‒173. 
F. Bömer, P. Ovidius Naso, Metamorphosen. Kommentar, Buch I‒III, Heidelberg 1969. 
F. Bömer, P. Ovidius Naso, Metamorphosen. Kommentar, Buch IV‒V, Heidelberg 1976. 
F. H. Bothe, Vindiciae Ovidianae sive Annotationes in P. Ovidii Nasonis Metamorphoseon 

libros XV. Accedunt Joannis Henrici Vossii viri celeberrimi lectiones et notae, 
Gottingae 1818. 

P. Burman, Sylloges epistolarum a viris illustribus scriptarum tomi quique, collecti et digesti 
per Petrum Burmannum, Tomus III, Leidae 1727. 

J. G. Capoferreus, Animadversionum in auctores classicos L. L. Liber I qui est in Ovidii 
Metamorphosin, Lipsiae 1659. 

G. Celato, Nasonis vincere decus. Da Ovidio a Claudiano: gli studi di Nicolaus Heinsius sugli 
auctores latini, Pisa 2023. 

S. Díez Reboso, Edición crítica y comentario textual del libro XI de las Metamorfosis de 
Ovidio, Diss. Huelva 2014. 

A. Draeger, Ovid als Sprachbildner, Aurich 1888. 
R. Ellis, A Commentary on Catullus, Oxford 21889. 
T. Faber, Epistolae, Pars altera, Salmurii 1665. 
P. Fàbregas Salis, “Heinsius’s Vnus meus and Oxoniensis (or Balliolensis)”, ExClass 23, 2019, 

65‒108. 
P. Fàbregas Salis, “A note on Ovid, Metamorphoses 11, 17”, MD 84, 2020, 197‒205. 
P. Fàbregas Salis, El libro X de las Metamorfosis de Ovidio: un comentario textual, Huelva 

2022. 
P. Fedeli/R. Dimundo/I. Ciccarelli, Properzio. Elegie, Libro IV, Nordhausen 2015. 
N. Heinsius, Adversariorum libri IV, curante Petro Burmanno Juniore, Harlingae 1742. 
N. Heinsius, Commentarius in P. Ovidii Nasonis Opera Omnia, cur. I. F. Fischerus, Lipsiae 

1758. 
E. B. Lease, “Neve and Neque with the Imperative and Subjunctive”, AJPh 34, 1913, 418‒436. 
Ae. Linse, De P. Ovidio Nasone vocabulorum inventore, Lipsiae 1891. 
P. Loewe, Über die Praepositionen A, DE, EX bei Ovid (Beilage zum Programm des königl. 

Gymnasiums zu Strehlen 197), Strehlen 1889. 
G. Luck, “Emendaturus, si licuisset, erat”, Euphrosyne 36, 2008, 9‒19. 
P. Maas, Textkritik, Leipzig 41960. 
J. C. McKeown, Ovid. Amores. Text, Prolegomena and Commentary. Volume III: A 

Commentary on Book Two, Leeds 1998. 



 
 

F. Munari, “Codici Heinsiani degli Amores”, SIFC 24, 1950, 161‒165. 
F. Neue/C. Wagener, Formenlehre der lateinischen Sprache. Erster Band: Das Substantivum, 

Leipzig 31902. 
R. Pichon, De sermone amatorio apud Latinos elegiarum scriptores, Paris 1902. 
H. Pinkster, The Oxford Latin Syntax. Vol. 1: The Simple Clause, Oxford 2015. 
H. Pinkster, The Oxford Latin Syntax. Vol. 2: The Complex Sentence and Discourse, Oxford 

2021. 
M. Platnauer, Latin Elegiac Verse, Cambridge 1951. 
A. Ramírez de Verger, Book VI of Ovid’s Metamorphoses: A Textual Commentary, 

Berlin/Boston 2021. 
A. Ramírez de Verger, “Textual criticism”, in: L. Rivero (coord.), The Metamorphoses of Ovid: 

35 Years of Research 1980–2014, Vol. I (Lustrum 62, 2020), Göttingen 2022, 58‒215. 
M. D. Reeve, “Heinsiusʼs Manuscripts of Ovid”, RhM 117, 1974, 133‒166. 
L. Rivero García, Book XIII of Ovid’s Metamorphoses: A Textual Commentary, Berlin/Boston 

2018. 
D. A. Slater, Towards a Text of the Metamorphosis of Ovid, Oxford 1927. 
L. Sniehotta, De vocum Graecarum apud poetas Latinos dactylicos ab Enni usque ad Ovidi 

tempora usu, Vratislaviae 1903. 
A. Suárez del Río, Edición crítica y comentario textual del libro III de las Metamorfosis de 

Ovidio, Diss. Huelva 2015. 
A. Szantyr, Lateinische Grammatik. Band II: Syntax und Stilistik, München 1972 (≈ 1965). 
W. S. Watt, “Ovidiana”, MH 52, 1995, 90‒107. 
W. S. Watt, “Notes on Ovid”, C&M 50, 1999, 167‒178. 
J. Wills, Repetition in Latin Poetry. Figures of Allusion, Oxford 1996. 
U. Winter, Die europäischen Handschriften der Bibliothek Diez. Zweiter Teil: Die Libri 

impressi cum notis manuscriptis der Bibliotheca Dieziana, Leipzig 1986. 
 
 
Pere Fàbregas Salis 
Departamento de Filología Clásica  
Universidad Complutense de Madrid  
Plaza Menéndez Pelayo s/n, 28040 Madrid, Spain  
E-Mail: pfabregasuni@gmail.com